First water cooled pc in two decades

[ThrashZone]

I like this one hehe

 

[A Computer Guy]

I like this one hehe
View attachment 333539

Looks like a lot of vitamin DD in there.

 

[A Computer Guy]

Something just seems to be right on the verge of not being capable enough. I would think that pump would be more than sufficient for this simple loop.

Can you describe this better? Is your pump making noise, or flow not fast, or temps not good? I was thinking the other day if I was to use the coolant you did I would probably cut it again with 50% distilled water depending on the viscosity of the coolant.

Congrats on the build that copper looks nice.

 

[67Elco]

Look at the Cinebench run above with the stats from HWinfo...not much headroom if any. Just imagine if this were not a 65w cpu.

 

[ThrashZone]

You can look at lots of reviews and HWlabs is right there at the top

I have a couple 280 gtx 45mm thick I believe designed for high speed fans
GTS is designed for low speed fans.

 

[ThrashZone]

Best PC radiator for water cooling in 2024 - 120mm to 480mm

One of the least glamorous, however, vitally important components to your liquid cooling loop is the radiator. The PC radiator is responsible

www.wepc.com

 

[A Computer Guy]

Look at the Cinebench run above with the stats from HWinfo...not much headroom if any. Just imagine if this were not a 65w cpu.

Ok so I see your air cooling test was performing better. I didn't see if you had a temp sensor in your loop. That might be helpful in comparing when changing things in your system. You might try flipping the 3 fans around in the back/side to push some more fresh air into the case or flip at least that 1 fan next to the rads. I know in my O11D if I cover that same space next to the rad it starves my top rad of cooling performance. Alternatively you could try flipping your rad fans to intake.

 

[ThrashZone]

Making record time with this...

View attachment 333180

I'm just noticing you have all your fans as exhaust ?
The three on the rad are and the back wall is as well.

Is there intake fans I can't see ?
If not I would switch the easiest ones to intake.

 

[ThrashZone]

Personally I prefer intake on rads the cases already have filters.

 

[67Elco]

The three fans on the back are intake...reverse blades. The one on the side is conventional exhaust and the three on the radiator are pushing air thru the rad and out the top of the case. The psu is mounted upside down and is exhausting thru the bottom.

Seems most vendors don't list the fpi of their radiators. Found it - The rad I have is 20 fpi and rated for 1200w. Specs lead one to believe it is a pretty decent radiator.

Can't go wider than my current rad (120mm) in this case so it is pointless to consider another radiator.

 

[BoggledBeagle]

Beginning to wonder if I could have made a better choice in radiators...

I do not know if you noticed but I tested a slim 280 rad in the overclock section of the forum. Yesterday evening I tested a thick 360 rad and the difference IS HUGE. For example at 6°C delta, slim 280 cools 170W and thick 360 rad 400W!

 

[A Computer Guy]

I do not know if you noticed but I tested a slim 280 rad in the overclock section of the forum. Yesterday evening I tested a thick 360 rad and the difference IS HUGE. For example at 6°C delta, slim 280 cools 170W and thick 360 rad 400W!

Is that at 100% fans speed though?

 

[lilhasselhoffer]

To be honest. That depends where in the world you drive. When i get to the highway i accelerate to about 180-200 km//h and break down when the way is not free. Accelerating back to my target when others get back to the right lane so i can overtake them. There are also stretches where the speed is limited to 120km/h or even lower. Just think about: in germany there are roundabout 10% of the highways free of any speed limitation. On a nice sunday morning at a good piece of highway one is able to speed up to >300km/h to get into the Ü300-club (club of drivers who drove +300KM/h).

I don't disagree but it is not that correct in general.

Terminology to remember:
Highway - high speeds, with lights
Freeway - high speeds, without lights

The problem that you are missing is stopping for those lights...which seems to be the issue. Bringing this back to the car analogy is not perfectly clean...but it's an analogy not a description.

If you'd like that, then maybe let's add 10 complex dimensions to the mix. The car has to be able to idle, disable under-utilized cylinders, it has to have big and little cylinders (think E versus P cores), and it also has to have a bunch of components strapped on that generate heat but aren't the engine...because modern CPUs have things like the memory controller and PCI-e lanes that generate heat as some unknown rate but are relatively a black box. Of course by the time you get to a 1:1 analog of the car to a modern CPU the explanation requires ten pages of text to say what I did...so the next step is for you to have your eyes glaze over and screech "TL;DR" because anything requiring more than three sentences is beyond the ability of most to comprehend...especially when the point was "usage case defines viability of designs, and there is a value to a large reservoir for certain specific cases and not others." Nuance....


And this explanation took a big paragraph...so 90% of people stopped giving a crap after seeing more than two lines of words or bullet points.

 

[lilhasselhoffer]

I need to put one of those little spinning wheel thingies in my loop so I can see at a glance the pump is working...

Has anyone that uses antifreeze ever really had to flush/change their system for any specific reason? Just wondering how the mix reacts over time in this scenario.

As a heads up, glycol is anti-freeze and a component of brake fluid. Most of the preservatives that are commercially available use glycol in some quantity.

As far as thermal properties, I think Titanrig does a good job of summarizing things here: Titanrig water cooling article It's worth a review...



Do note that glycol and the like are the primary components of these additives. Brake fluid...as discussed elsewhere...is glycol esther, glycol ether, and some other bits. The difference being brake fluid "wet" makes it bad but water with impurities only slightly depresses its thermal conductivity (and thus heat transfer rate). A very simple version of this is 0.61>>0.26...so water is about twice as good at conducting heat...visualized in the table on the Titanrig article above.

I had a rig using one of the commercially prepared fluids and a sealed custom loop system running for about 7 years without issue (or fluid replacement). I flushed the system as a part of the cleaning and re-pasting of the components...but at that time the only issue was the tubing going a bit hard and a bit cloudy. That was definitely UV degrading instead of the fluid, because whenever it flushed there wasn't any residue.

 

[gurusmi]

Terminology to remember:
Highway - high speeds, with lights
Freeway - high speeds, without lights

In germany it is mostly defined:
Autobahn = Highway: No Lights, no crosses, separated lanes (own and opposite), at least 2 lanes per direction
Landstrasse = Lights, Villages communities etc, crossing streets, mostly 1 lane per direction (might be a Freeway in the us)

And exactly that differences in the definition of roads is the reason why i wrote that one has to differ where he rides. Driving in the US (or whereever) on a highway is quite different to driving in germany.

 

[BoggledBeagle]

Is that at 100% fans speed though?

Yes, Arctic P120 and P140 fans at full speed (on one side of the rads, push).

 

[BoggledBeagle]

I do not know if you noticed but I tested a slim 280 rad in the overclock section of the forum. Yesterday evening I tested a thick 360 rad and the difference IS HUGE. For example at 6°C delta, slim 280 cools 170W and thick 360 rad 400W!

I made an error again, the difference is not huge, in fact, tha rads are pretty similar. I think I finally figured out how to calculate the results.

 

[A Computer Guy]

The three fans on the back are intake...reverse blades. The one on the side is conventional exhaust and the three on the radiator are pushing air thru the rad and out the top of the case. The psu is mounted upside down and is exhausting thru the bottom.

Seems most vendors don't list the fpi of their radiators. Found it - The rad I have is 20 fpi and rated for 1200w. Specs lead one to believe it is a pretty decent radiator.

Can't go wider than my current rad (120mm) in this case so it is pointless to consider another radiator.

Looking at your case can you fit another 360 slim rad in the rear/side compartment?

It seems you might be able to do something like this but it depends if you have enough room in the rear/side chamber.

For example in my 011D I had just enough room for the slim rad and fans. (of course the case panel had ventilation holes for this area of the case or this wouldn't work)

(edit)

Humm, maybe not. The I/O will be in the way for a 360 and the cross sections will block ports for a 240.
Perhaps if you flip it around and have the fans in the rear/side chamber instead you might be able to get away with another slim 360 in the inner chamber (with 2 fans in the outer chamber).
Looks like if you move the pump out a bit you can fit rad (ports bottom) in the inner chamber.

 

[Bobaganoosh]

I'm curious, what kind of thermal paste did you use? I skimmed the thread, but at 7 pages, I may have missed it.

I have usually found with water blocks that if my performance is not what I expect, it's commonly a bad mount. This has been (at least for me), a bit difficult with the water block as there's no spring mechanism or "bottom-out" type of scenario, so you just have to tighten very evenly with not too little or not too much pressure and an uneven mount is not uncommon at all, especially if you're installing with the case in the up-right position, which I always do on account of how large and heavy my case is (among other issues lol). So if your tubing allows for any kind of back-off of the block to clean and re-apply paste, I'd go with the best paste you can get and try to make sure the mount is level with the IHS as best you can.

While AMD CPUs have so much less power in them than Intel lately, I still find the paste thermal conductivity is extremely important because the thermal density is still quite high due to the concentrated die area under the IHS. The heat conducts through the thin IHS much easier than it spreads in it so having the best possible conduction path to the water block is still important. I've had the best results with Kingpin KPx, Kryonaut Extreme (the pink one), and Prolimatech PK-3 nano. The PK-3 nano is the longest-lasting paste of those though and if available it's usually pretty cheap so I love that stuff. the KPx is the easiest to buy (and easy to work with generally), but there are plenty of stories of pump-out after a year or so with that and the Kryo-extreme causing performance degradation (and needing a re-application). With hard-line tubing and not wanting to re-do it every year, the PK-3 is probably a good way to go. There are other good pastes too, but I can only speak for the ones I've tested personally and nothing else I've tried gets close to those.

 

[damric]

It's a work of art. Very nice indeed.

I am guessing your CPU has PBO enabled so the idle temperature is normal because they sit there and try to boost like crazy for no reason

Otherwise you pretty much max'd out what that case can do as far as radiators, pump, and reservoir. Anything more would require external radiator, and additional pump, and a larger reservoir to feed those pumps at full speed.

Speaking of which, large reservoirs are for maintaining net positive suction head. If your pump isn't cavitating, then you are fine with what you have. I run my dual D5s at full speed so I need a larger reservoir.

Don't be alarmed at your tank level dropping as tiny amounts of air is worked out. Just refill.

But yeah love the copper. It's like the tubes are acting as an additional radiator, plus they look awesome!

 

[67Elco]

I've given it some thought and perhaps I have been unfair in my comparison to my Noctua D15. It's true I expected slightly better thermals at the top with water, but I may well be chasing ghosts.

I'm going to back off the aggressive tune for a more sensible setting and re-run some benchmarks. I'll run them again with more sensible settings and post them for your honest reviews. Cinebench24 and AIDA64 are all I have currently...let me know if there are others which would be helpful.

The thermal paste used here is some old Arctic Alumina and I have some Noctua paste which I have never tried. No doubt there are better things available now, but I don't see how going down that rabbit hole will make a huge difference. The D5 pump is rated at 1100 L/h...I'll leave it for those with more experience to judge it's suitability in this application.

About to go back into bios and reset all to the same I used with the air cooling...basically just PBO with a negative 30 on all cores and nothing else.

 

[damric]

I've given it some thought and perhaps I have been unfair in my comparison to my Noctua D15. It's true I expected slightly better thermals at the top with water, but I may well be chasing ghosts.

I'm going to back off the aggressive tune for a more sensible setting and re-run some benchmarks. I'll run them again with more sensible settings and post them for your honest reviews. Cinebench24 and AIDA64 are all I have currently...let me know if there are others which would be helpful.

The thermal paste used here is some old Arctic Alumina and I have some Noctua paste which I have never tried. No doubt there are better things available now, but I don't see how going down that rabbit hole will make a huge difference. The D5 pump is rated at 1100 L/h...I'll leave it for those with more experience to judge it's suitability in this application.

About to go back into bios and reset all to the same I used with the air cooling...basically just PBO with a negative 30 on all cores and nothing else.

It's not like you will be running CB and AIDA all the time so you should really be fine with normal bursty workloads. I'm sure your coolant temp is just a couple degrees above ambient, if that.

If you really want to test the overall cooling capacity and true clock and power stability, try folding@home for a few days and take note of the trends of your CPU, GPU and coolant temperature (just be sure to finish your work units).

 

[67Elco]

Well, that was interesting. While I was in bios I changed the platform thermal limit to 85c from 95c...appears there was a about a 5c reduction in temps with no degraded performance. Perhaps I should re-run it on auto?

 

[ShrimpBrime]

Well, that was interesting. While I was in bios I changed the platform thermal limit to 85c from 95c...appears there was a about a 5c reduction in temps with no degraded performance. Perhaps I should re-run it on auto?

View attachment 333613

For sure. More tests, more information.
More repeated tests, better averages.

And the paste, find better paste. I second that. Alumina isn't great. Only 4w/mk, I would look for a min of 8.

 

[67Elco]

I'm going to pop into Starfield and get a screenshot of the cpu temps in game. When on air cooling I used to see an average of 65c.